R Notebook

1 . Setting up

if (!require(pacman))
  install.packages("pacman")
pacman::p_load(tibble, readr, dplyr, tidyr, lubridate, stringr, ggplot2, PBSmapping, animation)
pacman::p_load_gh("sjmgarnier/graphZoo")

2 . Loading and preparing data

if (!file.exists("data.csv")) {
  tmp <- read_delim("http://app.ntsb.gov/aviationquery/Download.ashx?type=csv", 
                    "|", na = c("NA", "", "N/A"))
  write_csv(tmp[, 1:(ncol(tmp) - 1)], "data.csv")
}
  
raw_dat <- read_csv("data.csv") %>%
  mutate(`Event Date` = mdy(`Event Date`)) %>%
  filter(year(`Event Date`) >= 1982, year(`Event Date`) < 2018,
         `Investigation Type` == "Accident", Country == "United States") %>%
  mutate(`Total Fatal Injuries` = ifelse(is.na(`Total Fatal Injuries`), 0, `Total Fatal Injuries`),
         `Total Serious Injuries` = ifelse(is.na(`Total Serious Injuries`), 0, `Total Serious Injuries`),
         `Total Minor Injuries` = ifelse(is.na(`Total Minor Injuries`), 0, `Total Minor Injuries`))

Parsed with column specification:
cols(
  .default = col_character(),
  Latitude = col_double(),
  Longitude = col_double(),
  `Number of Engines` = col_integer(),
  `Total Fatal Injuries` = col_integer(),
  `Total Serious Injuries` = col_integer(),
  `Total Minor Injuries` = col_integer(),
  `Total Uninjured` = col_integer()
)
See spec(...) for full column specifications.

raw_dat

3 . Time

##. Preparing data

byday_dat <- raw_dat %>%
  group_by(`Event Date`) %>%
  summarize(Events = n(),
            `Total Fatal Injuries` = sum(`Total Fatal Injuries`, na.rm = TRUE),
            `Total Serious Injuries` = sum(`Total Serious Injuries`, na.rm = TRUE),
            `Total Minor Injuries` = sum(`Total Minor Injuries`, na.rm = TRUE)) %>%
  complete(`Event Date` = seq.Date(ymd("1982-01-01"), ymd("2017-12-31"), by = 1),
           fill = list(Events = 0,
                       `Total Fatal Injuries` = 0, 
                       `Total Serious Injuries` = 0,
                       `Total Minor Injuries` = 0)) %>%
  mutate(Year = year(`Event Date`),
         Month = month(`Event Date`, label = TRUE),
         Week = week(`Event Date`),
         Day = wday(`Event Date`, label = TRUE, week_start = 1))
byday_dat

3.1 . Number of accidents per month since 1982

g <- byday_dat %>%
  group_by(Month_end = ceiling_date(`Event Date`, unit = "months")) %>%
  summarize(Events = sum(Events)) %>%
  ggplot(aes(x = Month_end, y = Events)) +
  geom_line() +
  geom_line(stat = "smooth", method = loess, se = FALSE, size = 2, color = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation accidents, 1982-2017", size = 8)
gSubtitle <- gz_title("per month, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.2 . Number of fatalities per month since 1982

g <- byday_dat %>%
  group_by(Month_end = ceiling_date(`Event Date`, unit = "months")) %>%
  summarize(`Total Fatal Injuries` = sum(`Total Fatal Injuries`)) %>%
  ggplot(aes(x = Month_end, y = `Total Fatal Injuries`)) +
  geom_line() +
  geom_line(stat = "smooth", method = loess, se = FALSE, size = 2, color = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation fatalities, 1982-2017", size = 8)
gSubtitle <- gz_title("per month, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.3 . Number of injuries per month since 1982

g <- byday_dat %>%
  group_by(Month_end = ceiling_date(`Event Date`, unit = "months")) %>%
  summarize(Injuries = sum(`Total Serious Injuries`) + sum(`Total Minor Injuries`)) %>%
  ggplot(aes(x = Month_end, y = Injuries)) +
  geom_line() +
  geom_line(stat = "smooth", method = loess, se = FALSE, size = 2, color = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation non-lethal injuries, 1982-2017", size = 8)
gSubtitle <- gz_title("per month, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.4 . Month with most accidents

g <- byday_dat %>%
  group_by(Month) %>%
  summarize(total = sum(Events)) %>%
  ggplot(aes(x = Month, y = total, group = 1)) +
  geom_bar(stat = "identity", color = "#DB4A36", fill = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation accidents, 1982-2017", size = 8)
gSubtitle <- gz_title("total by month, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.5 . Days with most accidents

g <- byday_dat %>%
  group_by(Day) %>%
  summarize(total = sum(Events)) %>%
  ggplot(aes(x = Day, y = total, group = 1)) +
  geom_bar(stat = "identity", color = "#DB4A36", fill = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation accidents, 1982-2017", size = 8)
gSubtitle <- gz_title("total by day, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.6 . Month with most fatalities

g <- byday_dat %>%
  group_by(Month) %>%
  summarize(total = sum(`Total Fatal Injuries`)) %>%
  ggplot(aes(x = Month, y = total, group = 1)) +
  geom_bar(stat = "identity", color = "#DB4A36", fill = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation fatalities, 1982-2017", size = 8)
gSubtitle <- gz_title("total by month, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.7 . Days with most fatalities

g <- byday_dat %>%
  group_by(Day) %>%
  summarize(total = sum(`Total Fatal Injuries`)) %>%
  ggplot(aes(x = Day, y = total, group = 1)) +
  geom_bar(stat = "identity", color = "#DB4A36", fill = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation fatalities, 1982-2017", size = 8)
gSubtitle <- gz_title("total by day, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.8 . Month with most injuries

g <- byday_dat %>%
  group_by(Month) %>%
  summarize(total = sum(`Total Serious Injuries`) + sum(`Total Minor Injuries`)) %>%
  ggplot(aes(x = Month, y = total, group = 1)) +
  geom_bar(stat = "identity", color = "#DB4A36", fill = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation non-lethal injuries, 1982-2017", size = 8)
gSubtitle <- gz_title("total by month, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

3.9 . Days with most injuries

g <- byday_dat %>%
  group_by(Day) %>%
  summarize(total = sum(`Total Serious Injuries`) + sum(`Total Minor Injuries`)) %>%
  ggplot(aes(x = Day, y = total, group = 1)) +
  geom_bar(stat = "identity", color = "#DB4A36", fill = "#DB4A36", alpha = 0.75) +
  ylim(0, NA) +
  xlab(NULL) + ylab(NULL) + 
  theme_graphzoo(base_size = 20)
gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                     r.txt = "DATA: NTSB",
                     font.size = 4)
gTitle <- gz_title("US civil aviation non-lethal injuries, 1982-2017", size = 8)
gSubtitle <- gz_title("total by day, within the USA, its territories and possessions, and in international waters", 
                      y = 0.9125, size = 5, fontface = "italic")
gz_combine(g, gTitle, gSubtitle, gBanner,
           top = c(0.9, 1, 1, 0.05), 
           bottom = c(0.0125, 0, 0, 0))

4 . Space

4.1 . Preparing data

loc_dat <- raw_dat %>%
  filter(year(`Event Date`) >= 2002, year(`Event Date`) < 2018,
         !is.na(Latitude), !is.na(Longitude)) %>%
  mutate(Date_round = ceiling_date(`Event Date`, unit = "3 days"),
         Longitude = ifelse(Longitude > 0, -Longitude, Longitude))
loc_dat

xlim <- c(-180, -20)
ylim <- c(10, 72)
world_dat <- map_data("world") %>%
  rename(X = long, Y = lat, PID = group, POS = order) %>%
  clipPolys(xlim = xlim, ylim = ylim, keepExtra = TRUE) %>%
  as_tibble()
world_dat

4.2 . Animated map

if (!file.exists("animation.mp4")) {
  gBackground <- gz_background()
  gBanner <- gz_banner(l.txt = "GRAPHZOO.TUMBLR.COM", 
                       r.txt = "DATA: NTSB",
                       font.size = 4)
  gTitle <- gz_title("Locations of US civil aviation accidents, 2002-2017", size = 8)
  gSubtitle <- gz_title("within the USA, its territories and possessions, and in international waters", 
                        y = 0.9125, size = 5, fontface = "italic")
  
  saveVideo({
    dates <- seq.Date(ymd("2002-01-01"), ymd("2017-12-31"), by = 3)
    for (i in seq(1:length(dates))) {
      g <- ggplot() +
        geom_polygon(data = world_dat, aes(x = X, y = Y, group = PID),
                     fill = "gray50", color = "white") +
        geom_point(data = filter(loc_dat, Date_round <= dates[i], Date_round >= (dates[i] - 60)), 
                   aes(x = Longitude, y = Latitude, alpha = 60 - as.numeric(dates[i] - Date_round)), 
                   pch = 21, fill = "#DB4A36", color = "black", size = 3) +
        annotate("text", x = -180, y = 11, hjust = 0, vjust = 0,
                 label = format(dates[i], format = "%b %Y")) +
        coord_map(xlim = xlim, ylim = ylim) +
        xlab(NULL) + ylab(NULL) + guides(alpha = FALSE) + 
        theme_graphzoo(base_size = 18) +
        theme(axis.title = element_blank(), axis.text = element_blank(), axis.ticks = element_blank(),
              panel.grid.major = element_blank(), panel.grid.minor = element_blank())
      
      gz_combine(gBackground, g, gTitle, gSubtitle, gBanner,
                 top = c(1, 0.9, 1, 1, 0.05),
                 bottom = c(0.05, 0.0125, 0.05, 0.05, 0))
    }
  }, ani.height = 480, ani.width = 720, interval = 1 / 30)
}

Animation